Three Dow Cycles That Beat The Test Of Time

What the Presidential, Decennial and 3.5-Year Rhythms Say About 2026 and 2027

Last Thursday we discussed the 34-year cycle which sets the big picture for the Dow: about 25 years up, 9 years down or flat, with the current up phase running to about 2034. But it only works after 1940. So, I went looking for shorter rhythms that also held before 1940. Three stand out in 140 years of quarterly Dow data, and they say something about 2026 and 2027.

How I tested them:

I took Dow closes from 1886 to November 2026, removed the long-term trend, and ran a frequency analysis to find which cycle lengths carry more swing than random data of the same smoothness would. Then I split the history in two, 1886 to 1956 and 1956 to 2026, and asked two questions: does the cycle show up in both halves, and does its timing carry over from one half to the other? A cycle that only fits the data it was found in is worth nothing.

1. The 4-year presidential cycle

The idea goes back to Yale Hirsch's Stock Trader's Almanac in the 1960s: markets tend to be weak in the second year of a presidential term and strong in the third, the year before the next election.

The shape is the same in both eras: the midterm year is the weakest of the four, and the pre-election year is strong. Since 1956 the pre-election year has been up 16 times out of 17. The usual explanation is political: administrations take the hard medicine early in the term and turn to stimulus before the next election.

2. The decennial cycle

Edgar Lawrence Smith described it in 1939: the Dow tends to follow a ten-year pattern by the last digit of the year. Of the three cycles, this one carried over best from one half of the history to the other.

Years ending in 5 were the best in both eras, and 2025 kept the record with a gain of about 13%. Years ending in 0 have been weak or flat in both. So far the DJIA is showing stronger correlation with the year 6 bullish decennial cycle:

The samples are small, however, five to nine years per digit, so read the pattern as a tilt, not a rule.

3. The 3.5-year cycle

Joseph Kitchin found a cycle of about 40 months in business inventories in 1923. It is the strongest short rhythm in the Dow: it shows up clearly in both halves of the history. Counting the major lows (the lowest close within 15 months on either side), they came a median 3.3 years apart before 1940 and 3.4 years apart since.

Recent lows: 2002, 2009, 2016, 2020 and October 2022. In long bull markets, like the 1990s, the dips that mark the cycle can be too shallow to count.

There is a catch. The cycle is real, but its timing drifts: a fixed 3.5-year wave fitted to one era did no better than a coin toss in the other. So it can't be projected from a formula. It has to be counted from the actual lows. From the October 2022 low, the usual spacing (2.5 to 4.4 years) puts the next low between spring 2025 and spring 2027. The April 2025 drop of about 16% may have been it.

What they say now

So far 2026 is up about 7% for the Dow, better than the average midterm year. Midterm years often bottom in the summer or fall and then rally into the pre-election year. If a pullback comes before year-end, the cycles would read it as the setup for 2027, not the start of something worse. Further out, 2030, a year ending in 0, has been a soft spot in the decennial pattern.

What didn't make the cut

A 6-year rhythm showed up only before 1956 and a 2-year rhythm only after. Neither carried over from one era to the other. The longest cycle in the data, about 37 years, is the 34-year cycle again.

Data: Dow Jones Industrial Average daily closes, 1886 to November 7, 2026; annual changes from year-end to year-end. Cycles describe history; they don't guarantee anything. For information only, not investment advice.